lychiyu/freqtrade_origin
1
0
Fork 0
mirror of https://github.com/freqtrade/freqtrade.git synced 2024-11-16 05:03:55 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity
7ae69a9cde
freqtrade_origin/freqtrade/data/converter/converter.py
Joe Schr 849413c5ea Changes scale from type int to float
2024-02-10 11:53:40 +01:00

614 lines
26 KiB
Python
Raw Blame History

"""
Functions to convert data from one format to another
"""
import logging
import time
from typing import Dict, List
import numpy as np
import pandas as pd
from pandas import DataFrame, to_datetime
from freqtrade.constants import (DEFAULT_DATAFRAME_COLUMNS, DEFAULT_ORDERFLOW_COLUMNS,
DEFAULT_TRADES_COLUMNS, Config)
from freqtrade.data.converter.trade_converter import trades_df_remove_duplicates
from freqtrade.enums import CandleType, TradingMode
logger = logging.getLogger(__name__)
def ohlcv_to_dataframe(ohlcv: list, timeframe: str, pair: str, *,
fill_missing: bool = True, drop_incomplete: bool = True) -> DataFrame:
"""
Converts a list with candle (OHLCV) data (in format returned by ccxt.fetch_ohlcv)
to a Dataframe
:param ohlcv: list with candle (OHLCV) data, as returned by exchange.async_get_candle_history
:param timeframe: timeframe (e.g. 5m). Used to fill up eventual missing data
:param pair: Pair this data is for (used to warn if fillup was necessary)
:param fill_missing: fill up missing candles with 0 candles
(see ohlcv_fill_up_missing_data for details)
:param drop_incomplete: Drop the last candle of the dataframe, assuming it's incomplete
:return: DataFrame
"""
logger.debug(f"Converting candle (OHLCV) data to dataframe for pair {pair}.")
cols = DEFAULT_DATAFRAME_COLUMNS
df = DataFrame(ohlcv, columns=cols)
df['date'] = to_datetime(df['date'], unit='ms', utc=True)
# Some exchanges return int values for Volume and even for OHLC.
# Convert them since TA-LIB indicators used in the strategy assume floats
# and fail with exception...
df = df.astype(dtype={'open': 'float', 'high': 'float', 'low': 'float', 'close': 'float',
'volume': 'float'})
return clean_ohlcv_dataframe(df, timeframe, pair,
fill_missing=fill_missing,
drop_incomplete=drop_incomplete)
def _init_dataframe_with_trades_columns(dataframe: DataFrame):
"""
Populates a dataframe with trades columns
:param dataframe: Dataframe to populate
"""
dataframe['trades'] = dataframe.apply(lambda _: [], axis=1)
dataframe['orderflow'] = dataframe.apply(lambda _: {}, axis=1)
dataframe['bid'] = np.nan
dataframe['ask'] = np.nan
dataframe['delta'] = np.nan
dataframe['min_delta'] = np.nan
dataframe['max_delta'] = np.nan
dataframe['total_trades'] = np.nan
dataframe['stacked_imbalances_bid'] = np.nan
dataframe['stacked_imbalances_ask'] = np.nan
def _convert_timeframe_to_pandas_frequency(timeframe: str):
# convert timeframe to format usable by pandas
from freqtrade.exchange import timeframe_to_minutes
timeframe_minutes = timeframe_to_minutes(timeframe)
timeframe_frequency = f'{timeframe_minutes}min'
return (timeframe_frequency, timeframe_minutes)
def _calculate_ohlcv_candle_start_and_end(df: DataFrame, timeframe: str):
timeframe_frequency, timeframe_minutes = _convert_timeframe_to_pandas_frequency(
timeframe)
# calculate ohlcv candle start and end
if df is not None and not df.empty:
df['datetime'] = pd.to_datetime(df['date'], unit='ms')
df['candle_start'] = df['datetime'].dt.floor(timeframe_frequency)
df['candle_end'] = df['candle_start'] + pd.Timedelta(timeframe_minutes)
df.drop(columns=['datetime'], inplace=True)
def populate_dataframe_with_trades(config: Config,
dataframe: DataFrame,
trades: DataFrame,
*,
pair: str) -> DataFrame:
"""
Populates a dataframe with trades
:param dataframe: Dataframe to populate
:param trades: Trades to populate with
:return: Dataframe with trades populated
"""
config_orderflow = config['orderflow']
timeframe = config['timeframe']
# create columns for trades
_init_dataframe_with_trades_columns(dataframe)
df = dataframe.copy()
try:
start_time = time.time()
# calculate ohlcv candle start and end
_calculate_ohlcv_candle_start_and_end(df, timeframe)
_calculate_ohlcv_candle_start_and_end(trades, timeframe)
# slice of trades that are before current ohlcv candles to make groupby faster
trades = trades.loc[trades.candle_start >= df.candle_start[0]]
trades.reset_index(inplace=True, drop=True)
# group trades by candle start
trades_grouped_by_candle_start = trades.groupby(
'candle_start', group_keys=False)
# repair 'date' datetime type (otherwise crashes on each compare)
if "date" in dataframe.columns:
dataframe['date'] = pd.to_datetime(dataframe['date'])
for candle_start in trades_grouped_by_candle_start.groups:
trades_grouped_df = trades[candle_start == trades['candle_start']]
is_between = (candle_start == df['candle_start'])
if np.any(is_between == True): # noqa: E712
(timeframe_frequency, timeframe_minutes) = _convert_timeframe_to_pandas_frequency(
timeframe)
candle_next = candle_start + \
pd.Timedelta(minutes=timeframe_minutes)
# skip if there are no trades at next candle
# because that this candle isn't finished yet
if candle_next not in trades_grouped_by_candle_start.groups:
logger.warning(
f"candle at {candle_start} with {len(trades_grouped_df)} trades might be unfinished, because no finished trades at {candle_next}") # noqa
# add trades to each candle
df.loc[is_between, 'trades'] = df.loc[is_between,
'trades'].apply(lambda _: trades_grouped_df)
# calculate orderflow for each candle
df.loc[is_between, 'orderflow'] = df.loc[is_between, 'orderflow'].apply(
lambda _: trades_to_volumeprofile_with_total_delta_bid_ask(
pd.DataFrame(trades_grouped_df),
scale=config_orderflow['scale']))
# calculate imbalances for each candle's orderflow
df.loc[is_between, 'imbalances'] = df.loc[is_between, 'orderflow'].apply(
lambda x: trades_orderflow_to_imbalances(x,
imbalance_ratio=config_orderflow['imbalance_ratio'], # noqa: E501
imbalance_volume=config_orderflow['imbalance_volume'])) # noqa: E501
df.loc[is_between, 'stacked_imbalances_bid'] = df.loc[is_between,
'imbalances'].apply(
lambda x: stacked_imbalance_bid(x, # noqa: E501
stacked_imbalance_range=config_orderflow['stacked_imbalance_range'])) # noqa: E501
df.loc[is_between, 'stacked_imbalances_ask'] = df.loc[is_between,
'imbalances'].apply(
lambda x: stacked_imbalance_ask(x, # noqa: E501
stacked_imbalance_range=config_orderflow['stacked_imbalance_range'])) # noqa: E501
buy = df.loc[is_between, 'bid'].apply(lambda _: np.where(
trades_grouped_df['side'].str.contains('buy'), 0, trades_grouped_df['amount']))
sell = df.loc[is_between, 'ask'].apply(lambda _: np.where(
trades_grouped_df['side'].str.contains('sell'), 0, trades_grouped_df['amount']))
deltas_per_trade = sell - buy
min_delta = 0
max_delta = 0
delta = 0
for deltas in deltas_per_trade:
for d in deltas:
delta += d
if delta > max_delta:
max_delta = delta
if delta < min_delta:
min_delta = delta
df.loc[is_between, 'max_delta'] = max_delta
df.loc[is_between, 'min_delta'] = min_delta
df.loc[is_between, 'bid'] = np.where(trades_grouped_df['side'].str.contains(
'buy'), 0, trades_grouped_df['amount']).sum()
df.loc[is_between, 'ask'] = np.where(trades_grouped_df['side'].str.contains(
'sell'), 0, trades_grouped_df['amount']).sum()
df.loc[is_between, 'delta'] = df.loc[is_between,
'ask'] - df.loc[is_between, 'bid']
min_delta = np.min(deltas_per_trade)
max_delta = np.max(deltas_per_trade)
df.loc[is_between, 'total_trades'] = len(trades_grouped_df)
# copy to avoid memory leaks
dataframe.loc[is_between] = df.loc[is_between].copy()
else:
logger.debug(
f"Found NO candles for trades starting with {candle_start}")
logger.debug(
f"trades.groups_keys in {time.time() - start_time} seconds")
logger.debug(
f"trades.singleton_iterate in {time.time() - start_time} seconds")
except Exception as e:
logger.exception("Error populating dataframe with trades:", e)
return dataframe
def public_trades_to_dataframe(trades: List,
pair: str,
) -> DataFrame:
"""
Converts a list with candle (TRADES) data (in format returned by ccxt.fetch_trades)
to a Dataframe
:param trades: list with candle (TRADES) data, as returned by exchange.async_get_candle_history
:param timeframe: timeframe (e.g. 5m). Used to fill up eventual missing data
:param pair: Pair this data is for (used to warn if fillup was necessary)
:param fill_missing: fill up missing candles with 0 candles
(see trades_fill_up_missing_data for details)
:param drop_incomplete: Drop the last candle of the dataframe, assuming it's incomplete
:return: DataFrame
"""
logger.debug(
f"Converting candle (TRADES) data to dataframe for pair {pair}.")
cols = DEFAULT_TRADES_COLUMNS
df = DataFrame(trades, columns=cols)
df['date'] = pd.to_datetime(
df['timestamp'], unit='ms', utc=True)
# Some exchanges return int values for Volume and even for OHLC.
# Convert them since TA-LIB indicators used in the strategy assume floats
# and fail with exception...
df = df.astype(dtype={'amount': 'float', 'cost': 'float',
'price': 'float'})
return df
def trades_to_volumeprofile_with_total_delta_bid_ask(trades: DataFrame, scale: float):
"""
:param trades: dataframe
:param scale: scale aka bin size e.g. 0.5
:return: trades binned to levels according to scale aka orderflow
"""
df = pd.DataFrame([], columns=DEFAULT_ORDERFLOW_COLUMNS)
# create bid, ask where side is sell or buy
df['bid_amount'] = np.where(
trades['side'].str.contains('buy'), 0, trades['amount'])
df['ask_amount'] = np.where(
trades['side'].str.contains('sell'), 0, trades['amount'])
df['bid'] = np.where(
trades['side'].str.contains('buy'), 0, 1)
df['ask'] = np.where(
trades['side'].str.contains('sell'), 0, 1)
# round the prices to the nearest multiple of the scale
df['price'] = ((trades['price'] / scale).round()
* scale).astype('float64').values
if df.empty:
df['total'] = np.nan
df['delta'] = np.nan
return df
df['delta'] = df['ask_amount'] - df['bid_amount']
df['total_volume'] = df['ask_amount'] + df['bid_amount']
df['total_trades'] = df['ask'] + df['bid']
# group to bins aka apply scale
df = df.groupby('price').sum(numeric_only=True)
return df
def trades_orderflow_to_imbalances(df: DataFrame, imbalance_ratio: int, imbalance_volume: int):
"""
:param df: dataframes with bid and ask
:param imbalance_ratio: imbalance_ratio e.g. 300
:param imbalance_volume: imbalance volume e.g. 3)
:return: dataframe with bid and ask imbalance
"""
bid = df.bid
ask = df.ask.shift(-1)
bid_imbalance = (bid / ask) > (imbalance_ratio / 100)
# overwrite bid_imbalance with False if volume is not big enough
bid_imbalance_filtered = np.where(
df.total_volume < imbalance_volume, False, bid_imbalance)
ask_imbalance = (ask / bid) > (imbalance_ratio / 100)
# overwrite ask_imbalance with False if volume is not big enough
ask_imbalance_filtered = np.where(
df.total_volume < imbalance_volume, False, ask_imbalance)
dataframe = DataFrame(
{"bid_imbalance": bid_imbalance_filtered,
"ask_imbalance": ask_imbalance_filtered},
index=df.index,
)
return dataframe
def stacked_imbalance(df: DataFrame,
label: str = "bid",
stacked_imbalance_range: int = 3,
should_reverse: bool = False):
"""
y * (y.groupby((y != y.shift()).cumsum()).cumcount() + 1)
https://stackoverflow.com/questions/27626542/counting-consecutive-positive-values-in-python-pandas-array
"""
imbalance = df[f'{label}_imbalance']
int_series = pd.Series(np.where(imbalance, 1, 0))
stacked = int_series * \
(int_series.groupby((int_series != int_series.shift()).cumsum()).cumcount() + 1)
max_stacked_imbalance_idx = stacked.index[stacked >=
stacked_imbalance_range]
stacked_imbalance_price = np.nan
if not max_stacked_imbalance_idx.empty:
idx = max_stacked_imbalance_idx[0] if not should_reverse else np.flipud(
max_stacked_imbalance_idx)[0]
stacked_imbalance_price = imbalance.index[idx]
return stacked_imbalance_price
def stacked_imbalance_bid(df: DataFrame, stacked_imbalance_range: int = 3):
return stacked_imbalance(df, 'bid', stacked_imbalance_range)
def stacked_imbalance_ask(df: DataFrame, stacked_imbalance_range: int = 3):
return stacked_imbalance(df, 'ask', stacked_imbalance_range, should_reverse=True)
def orderflow_to_volume_profile(df: DataFrame):
"""
:param orderflow: dataframe
:return: volume profile dataframe
"""
bid = df.groupby('level').bid.sum()
ask = df.groupby('level').ask.sum()
df.groupby('level')['level'].sum()
delta = df.groupby('level').ask.sum() - df.groupby('level').bid.sum()
df = pd.DataFrame({'bid': bid, 'ask': ask, 'delta': delta})
return df
def clean_ohlcv_dataframe(data: DataFrame, timeframe: str, pair: str, *,
fill_missing: bool, drop_incomplete: bool) -> DataFrame:
"""
Cleanse a OHLCV dataframe by
* Grouping it by date (removes duplicate tics)
* dropping last candles if requested
* Filling up missing data (if requested)
:param data: DataFrame containing candle (OHLCV) data.
:param timeframe: timeframe (e.g. 5m). Used to fill up eventual missing data
:param pair: Pair this data is for (used to warn if fillup was necessary)
:param fill_missing: fill up missing candles with 0 candles
(see ohlcv_fill_up_missing_data for details)
:param drop_incomplete: Drop the last candle of the dataframe, assuming it's incomplete
:return: DataFrame
"""
# group by index and aggregate results to eliminate duplicate ticks
data = data.groupby(by='date', as_index=False, sort=True).agg({
'open': 'first',
'high': 'max',
'low': 'min',
'close': 'last',
'volume': 'max',
})
# eliminate partial candle
if drop_incomplete:
data.drop(data.tail(1).index, inplace=True)
logger.debug('Dropping last candle')
if fill_missing:
return ohlcv_fill_up_missing_data(data, timeframe, pair)
else:
return data
def warn_of_tick_duplicates(data: DataFrame, pair: str) -> None:
no_dupes_colunms = ['id', 'timestamp', 'datetime']
for col in no_dupes_colunms:
if col in data.columns and data[col].duplicated().any():
sum = data[col].duplicated().sum()
message = f'{sum} duplicated ticks for {pair} in {col} detected.'
if col == 'id':
logger.warning(message)
else:
logger.debug(message)
def clean_duplicate_trades(trades: DataFrame, timeframe: str, pair: str, *,
fill_missing: bool, drop_incomplete: bool) -> DataFrame:
"""
Cleanse a TRADES dataframe by
* Grouping it by date (removes duplicate tics)
* dropping last candles if requested
* Filling up missing data (if requested)
:param data: DataFrame containing candle (TRADES) data.
:param timeframe: timeframe (e.g. 5m). Used to fill up eventual missing data
:param pair: Pair this data is for (used to warn if fillup was necessary)
:param fill_missing: fill up missing candles with 0 candles
(see trades_fill_up_missing_data for details)
:param drop_incomplete: Drop the last candle of the dataframe, assuming it's incomplete
:return: DataFrame
"""
# group by index and aggregate results to eliminate duplicate ticks
# check if data has duplicate ticks
logger.debug(f"Clean duplicated ticks from Trades data {pair}")
df = pd.DataFrame(trades_df_remove_duplicates(
trades), columns=trades.columns)
return df
def drop_incomplete_and_fill_missing_trades(data: DataFrame, timeframe: str, pair: str, *,
fill_missing: bool, drop_incomplete: bool) -> DataFrame:
# eliminate partial candle
if drop_incomplete:
# TODO: this is not correct, as it drops the last trade only
# but we need to drop the last candle until closed
pass
data.drop(data.tail(1).index, inplace=True)
logger.debug('Dropping last trade')
return data
def ohlcv_fill_up_missing_data(dataframe: DataFrame, timeframe: str, pair: str) -> DataFrame:
"""
Fills up missing data with 0 volume rows,
using the previous close as price for "open", "high" "low" and "close", volume is set to 0
"""
from freqtrade.exchange import timeframe_to_resample_freq
ohlcv_dict = {
'open': 'first',
'high': 'max',
'low': 'min',
'close': 'last',
'volume': 'sum'
}
resample_interval = timeframe_to_resample_freq(timeframe)
# Resample to create "NAN" values
df = dataframe.resample(resample_interval, on='date').agg(ohlcv_dict)
# Forwardfill close for missing columns
df['close'] = df['close'].ffill()
# Use close for "open, high, low"
df.loc[:, ['open', 'high', 'low']] = df[['open', 'high', 'low']].fillna(
value={'open': df['close'],
'high': df['close'],
'low': df['close'],
})
df.reset_index(inplace=True)
len_before = len(dataframe)
len_after = len(df)
pct_missing = (len_after - len_before) / len_before if len_before > 0 else 0
if len_before != len_after:
message = (f"Missing data fillup for {pair}, {timeframe}: "
f"before: {len_before} - after: {len_after} - {pct_missing:.2%}")
if pct_missing > 0.01:
logger.info(message)
else:
# Don't be verbose if only a small amount is missing
logger.debug(message)
return df
def trim_dataframe(df: DataFrame, timerange, *, df_date_col: str = 'date',
startup_candles: int = 0) -> DataFrame:
"""
Trim dataframe based on given timerange
:param df: Dataframe to trim
:param timerange: timerange (use start and end date if available)
:param df_date_col: Column in the dataframe to use as Date column
:param startup_candles: When not 0, is used instead the timerange start date
:return: trimmed dataframe
"""
if startup_candles:
# Trim candles instead of timeframe in case of given startup_candle count
df = df.iloc[startup_candles:, :]
else:
if timerange.starttype == 'date':
df = df.loc[df[df_date_col] >= timerange.startdt, :]
if timerange.stoptype == 'date':
df = df.loc[df[df_date_col] <= timerange.stopdt, :]
return df
def trim_dataframes(preprocessed: Dict[str, DataFrame], timerange,
startup_candles: int) -> Dict[str, DataFrame]:
"""
Trim startup period from analyzed dataframes
:param preprocessed: Dict of pair: dataframe
:param timerange: timerange (use start and end date if available)
:param startup_candles: Startup-candles that should be removed
:return: Dict of trimmed dataframes
"""
processed: Dict[str, DataFrame] = {}
for pair, df in preprocessed.items():
trimed_df = trim_dataframe(df, timerange, startup_candles=startup_candles)
if not trimed_df.empty:
processed[pair] = trimed_df
else:
logger.warning(f'{pair} has no data left after adjusting for startup candles, '
f'skipping.')
return processed
def order_book_to_dataframe(bids: list, asks: list) -> DataFrame:
"""
TODO: This should get a dedicated test
Gets order book list, returns dataframe with below format per suggested by creslin
-------------------------------------------------------------------
b_sum b_size bids asks a_size a_sum
-------------------------------------------------------------------
"""
cols = ['bids', 'b_size']
bids_frame = DataFrame(bids, columns=cols)
# add cumulative sum column
bids_frame['b_sum'] = bids_frame['b_size'].cumsum()
cols2 = ['asks', 'a_size']
asks_frame = DataFrame(asks, columns=cols2)
# add cumulative sum column
asks_frame['a_sum'] = asks_frame['a_size'].cumsum()
frame = pd.concat([bids_frame['b_sum'], bids_frame['b_size'], bids_frame['bids'],
asks_frame['asks'], asks_frame['a_size'], asks_frame['a_sum']], axis=1,
keys=['b_sum', 'b_size', 'bids', 'asks', 'a_size', 'a_sum'])
# logger.info('order book %s', frame )
return frame
def convert_ohlcv_format(
config: Config,
convert_from: str,
convert_to: str,
erase: bool,
):
"""
Convert OHLCV from one format to another
:param config: Config dictionary
:param convert_from: Source format
:param convert_to: Target format
:param erase: Erase source data (does not apply if source and target format are identical)
"""
from freqtrade.data.history.idatahandler import get_datahandler
src = get_datahandler(config['datadir'], convert_from)
trg = get_datahandler(config['datadir'], convert_to)
timeframes = config.get('timeframes', [config.get('timeframe')])
logger.info(f"Converting candle (OHLCV) for timeframe {timeframes}")
candle_types = [CandleType.from_string(ct) for ct in config.get('candle_types', [
c.value for c in CandleType])]
logger.info(candle_types)
paircombs = src.ohlcv_get_available_data(config['datadir'], TradingMode.SPOT)
paircombs.extend(src.ohlcv_get_available_data(config['datadir'], TradingMode.FUTURES))
if 'pairs' in config:
# Filter pairs
paircombs = [comb for comb in paircombs if comb[0] in config['pairs']]
if 'timeframes' in config:
paircombs = [comb for comb in paircombs if comb[1] in config['timeframes']]
paircombs = [comb for comb in paircombs if comb[2] in candle_types]
paircombs = sorted(paircombs, key=lambda x: (x[0], x[1], x[2].value))
formatted_paircombs = '\n'.join([f"{pair}, {timeframe}, {candle_type}"
for pair, timeframe, candle_type in paircombs])
logger.info(f"Converting candle (OHLCV) data for the following pair combinations:\n"
f"{formatted_paircombs}")
for pair, timeframe, candle_type in paircombs:
data = src.ohlcv_load(pair=pair, timeframe=timeframe,
timerange=None,
fill_missing=False,
drop_incomplete=False,
startup_candles=0,
candle_type=candle_type)
logger.info(f"Converting {len(data)} {timeframe} {candle_type} candles for {pair}")
if len(data) > 0:
trg.ohlcv_store(
pair=pair,
timeframe=timeframe,
data=data,
candle_type=candle_type
)
if erase and convert_from != convert_to:
logger.info(f"Deleting source data for {pair} / {timeframe}")
src.ohlcv_purge(pair=pair, timeframe=timeframe, candle_type=candle_type)
def reduce_dataframe_footprint(df: DataFrame) -> DataFrame:
"""
Ensure all values are float32 in the incoming dataframe.
:param df: Dataframe to be converted to float/int 32s
:return: Dataframe converted to float/int 32s
"""
logger.debug(f"Memory usage of dataframe is "
f"{df.memory_usage().sum() / 1024**2:.2f} MB")
df_dtypes = df.dtypes
for column, dtype in df_dtypes.items():
if column in ['open', 'high', 'low', 'close', 'volume']:
continue
if dtype == np.float64:
df_dtypes[column] = np.float32
elif dtype == np.int64:
df_dtypes[column] = np.int32
df = df.astype(df_dtypes)
logger.debug(f"Memory usage after optimization is: "
f"{df.memory_usage().sum() / 1024**2:.2f} MB")
return df
Reference in New Issue View Git Blame Copy Permalink